Deletion of Men1 and somatostatin induces hypergastrinemia and gastric carcinoids

Abstract

Background: Gastric carcinoids are slow growing neuroendocrine tumours arising from enterochromaffin-like (ECL) cells in the corpus of stomach. Although most of these tumours arise in the setting of gastric atrophy and hypergastrinemia, it is not understood what genetic background predisposes development of these ECL derived tumours. Moreover, diffuse microcarcinoids in the mucosa can lead to a field effect and limit successful endoscopic removal.

Objective: To define the genetic background that creates a permissive environment for gastric carcinoids using transgenic mouse lines.

Design: The multiple endocrine neoplasia 1 gene locus (Men1) was deleted using Cre recombinase expressed from the Villin promoter (Villin-Cre) and was placed on a somatostatin null genetic background. These transgenic mice received omeprazole-laced chow for 6 months. The direct effect of gastrin and the gastrin receptor antagonist YM022 on expression and phosphorylation of the cyclin inhibitor p27^Kip1 was tested on the human human gastric adenocarcinoma cell line stably expressing CCKBR (AGSE) and mouse small intestinal neuroendocrine carcinoma (STC)-1 cell lines.

Results: The combination of conditional Men1 deletion in the absence of somatostatin led to the development of gastric carcinoids within 2 years. Suppression of acid secretion by omeprazole accelerated the timeline of carcinoid development to 6 months in the absence of significant parietal cell atrophy. Carcinoids were associated with hypergastrinemia, and correlated with increased Cckbr expression and nuclear export of p27^Kip1 both in vivo and in gastrin-treated cell lines. Loss of p27^Kip1 was also observed in human gastric carcinoids arising in the setting of atrophic gastritis.

Conclusions: Gastric carcinoids require threshold levels of hypergastrinemia, which modulates p27^Kip1 cellular location and stability.

Keywords: ACID SECRETION; CANCER GENETICS; ENDOCRINE TUMOURS; GASTRIN; GASTRIN RECEPTORS.

Figure 1. Gastric Carcinoid in Men1^ΔIEC; Sst^−/− mice

Stomachs from 23-month-old WT, Men1^ΔIEC, Sst^−/−, mice, showing polyps (arrows) in the corpi of two different Men1^ΔIEC; Sst^−/− mice (A). H&E staining of WT (B), Men1^ΔIEC (C), Sst^−/− (D), and Men1^ΔIEC; Sst^−/− mice (E, F) showing corpus hyperplasia and submucosal lesions in the latter. (G). DAB Immunohistochemistry for HDC of the area shown in F. (H). Immunofluorescent staining for Ki-67 and HDC in 23-month-old Men1^ΔIEC; Sst^−/− mice. (I). Stomachs from 8–9 month old untreated (−OM) and omeprazole treated (+OM) Men1^ΔIEC; Sst^−/− mice showing polyps (dotted lines) in the corpi. H&E staining of corpi of untreated (J) and omeprazole-treated Men1^ΔIEC; Sst^−/− mice (K), showing hyperplasia and classical “salt and pepper nuclei” (arrows) in the latter. (L, M). DAB Immunohistochemistry for HDC in omeprazole-treated Men1^ΔIEC; Sst^−/− mice. Scale bars: 100 µm.

Figure 2. Markedly high gastrin levels in omeprazole treated Men1^ΔIEC; Sst^−/− mice

Immunofluorescent staining for gastrin in the antra of untreated WT (A), Men1^ΔIEC (B), Sst^−/− (C), and Men1^ΔIEC; Sst^−/− mice (D), and mice treated with omeprazole for 6 months (E–H). (I). Antral gastrin content (nmol) from mice fasted for 16 hours normalized to tissue weight (n = 5–7). (J). Antral Gastrin mRNA expression measured by RT-qPCR normalized to Hprt mRNA (n = 5–8). (K). EIA measurement of gastrin concentration in circulating plasma from mice fasted for 16 h (n = 6–8). Shown are the Means ± SEM. * p< 0.05, ** p< 0.01, *** p< 0.001, **** p< 0.0001. Scale bars: 100 µm.

Figure 3. Increased Cckbr expression in Men1^ΔIEC; Sst^−/− mice treated with omeprazole

(A). Heat-map of transcripts generated from microarray analyses of RNA extracted from polyps and adjacent visibly normal areas of Men1^ΔIEC; Sst^−/− mice (n=2). (B). Quantitation of Cckbr expression in treated and untreated WT, Men1^ΔIEC, Sst^−/−, and Men1^ΔIEC; Sst^−/− mice, expressed as a ratio of mRNA from untreated WT controls, after normalization to Hprt. Shown are the Means ± SEM (n=5–8). (C). Representative immunoblot showing Cckbr protein levels (GAPDH used as loading control) in corpi of WT, Men1^ΔIEC, Sst^−/− mice and polyps isolated from Men1^ΔIEC; Sst^−/− mice. (D). Quantitation of integrated band intensities analyzed using LICOR Odyssey software, expressed as a ratio to untreated WT controls, after normalization to GAPDH. Shown are the Means ± SEM (n=5–8). (E). Representative image of Cckbr expression determined by DAB staining in untreated and omeprazole-treated Men1^ΔIEC; Sst^−/− mice. * p< 0.05, ** p< 0.01, *** p< 0.001. Scale bars: 100 µm.

Figure 4. Carcinoids are associated with altered p27^Kip1 expression

(A) Representative blot showing total p27^Kip1 expression in cellular lysates from corpi of untreated, or omeprazole-treated WT, Men1^ΔIEC, Sst^−/−, and polyps isolated from Men1^ΔIEC; Sst^−/− mice (β-Actin loading control). Representative blots showing total p27^Kip1, P-p27^Thr187 and Skp2 protein expressions in nuclear fraction (SP1 loading control); P-p27^Ser10 and P-p27^Thr198 in cytoplasmic fraction (GAPDH loading control). Nuclear and cytoplasmic fractions were prepared as described earlier and fraction purity was assessed using appropriate markers (nuclear fraction showed negligible GAPDH expression, and cytoplasmic fraction had negligible SP1 levels). (B). Quantitation of p27^Kip1 expression in in nuclear and cytoplasmic fractions, expressed as integrated band intensities normalized to appropriate loading controls (n=4–5). Letters (a–e) denote significant differences in nuclear and cytoplasmic p27^Kip1 expressions compared to that of respective cellular fractions in untreated WT controls. Bars with different letters are significantly different from each other. Asterisk symbols denote significant differences in total cellular p27^Kip1 expressions compared to that in untreated WT controls. Bars with different asterisk symbols are significantly different from each other. Quantitation of P-p27^Thr187 expression in nuclear fraction (C), and P-p27^Thr198 expression in cytoplasmic fraction (D) expressed as ratio of total p27^Kip1, after normalization to SP1, and GAPDH loading controls, respectively (n=4–6). Shown are the Means ± SEM. * p< 0.05, ** p< 0.01, *** p< 0.001. (E). Immunofluorescent staining for P-p27^Thr187, P-p27^Thr198 (arrows), and Hdc in the corpi of untreated or omeprazole-treated Men1^ΔIEC; Sst^−/− mice. Scale bars: 50 µm.

Figure 5. Gastrin induces CCKBR expression and nuclear export of p27^Kip1 in human AGSE cells

(A). Immunofluorescent staining of CCKBR and total p27^Kip1 (p27) in AGSE cells treated with or without 20 nM gastrin, in the presence and absence of 10 nM YM022 for 24 hours, and after removal of gastrin from the culture media. (B). Representative blot showing total p27^Kip1 in nuclear and cytoplasmic fractions of AGSE cells treated with or without 20 nM gastrin, in the presence and absence of 10 nM YM022 for 24 hours. (C). Quantitation of total p27^Kip1 expression in nuclear and cytoplasmic fractions of AGSE cells treated with or without gastrin and YM022. Shown are the Means ± SEM (n = 3 experiments). * p< 0.05. Letters (a, b) denote significant differences in nuclear and cytoplasmic p27^Kip1 expressions compared to respective cellular fractions in untreated cells (without gastrin or YM022). Asterisk denotes significant differences in total cellular p27^Kip1 expressions compared to that in untreated cells. (D). Representative blot showing total p27^Kip1 expression in nuclear and cytoplasmic fractions of AGSE cells treated without or with 1, 10, and 20 nM gastrin. Scale bars: 20 µm.

Figure 6. Increased CCKBR expression and loss of p27^Kip1 in human gastric carcinoids

DAB Immunohistochemistry for HDC, CCKBR, and p27^Kip1 in normal human stomach (A) and archived surgical specimens from gastric carcinoid patients (B). Scale bars: 20 µm.

Figure 7

Schematic representation of proposed mechanism for ECL cell hyperplasia and carcinoid formation.